1 of 28 9/15/2016 1:16 PM

Transcription

1 1 of 28 9/15/2016 1:16 PM

2 2 of 28 9/15/2016 1:16 PM

3 3 of 28 9/15/2016 1:16 PM

4 objects(package:psych).first < function(library(psych)) help(read.table) #or?read.table #another way of asking for help apropos("read") #returns all available functions with that term in their name. RSiteSearch("read") #opens a webbrowser and searches voluminous files 4 of 28 9/15/2016 1:16 PM

5 apropos(table) #lists all the commands that have the word "table" in them apropos(table) [1]"ftable" "model.tables" "pairwise.table" "print.ftable" "r2dtable" [6]"read.ftable" "write.ftable" ". C mtable" ". C summary.table" ". C table" [11]"as.data.frame.table" "as.table" "as.table.default" "is.table" "margin.table" [16]"print.summary.table" "print.table" "prop.table" "read.table" "read.table.url" [21]"summary.table" "table" "write.table" "write.table0" 5 of 28 9/15/2016 1:16 PM

6 6 of 28 9/15/2016 1:16 PM

7 datafilename < " project.org/r/datasets/finkel.sav" #remote file datafilename < "/Users/bill/Library/Favorites/R.tutorial/datasets/finkel.sav" #local file eli.data < read.spss(datafilename, use.value.labels=true, to.data.frame=true) #works for local but not remote? seems to be a problem in uploading to the server save(object,file="local name") #save an object (e.g., a correlation matrix) for later analysis load(file) #gets the object (e.g., the correlation matrix back) load(url(" project.org/r/datasets/big5r.txt")) #get the correlation matrix ls() #show the variables in the workspace datafilename < " project.org/r/datasets/maps.mixx.epi.bfi.data" person.data < read.table(datafilename,header=true) #read the data file names(person.data) #list the names of the variables attach(person.data) #make the separate variable available always do detach when finished. #The with construct is better. epi < cbind(epie,epis,epiimp,epilie,epineur) #form a new variable "epi" epi.df < data.frame(epi) #actually, more useful to treat this variable as a data frame bfi.df < data.frame(cbind(bfext,bfneur,bfagree,bfcon,bfopen)) #create bfi as a data frame as well detach(person.data) # very important to detach after an attach #alternatively: with(person.data,{ epi < cbind(epie,epis,epiimp,epilie,epineur) #form a new variable "epi" epi.df < data.frame(epi) #actually, more useful to treat this variable as a data frame bfi.df < data.frame(cbind(bfext,bfneur,bfagree,bfcon,bfopen)) #create bfi as a data frame as well epi.df < data.frame(epi) #actually, more useful to treat this variable as a data frame bfi.df < data.frame(cbind(bfext,bfneur,bfagree,bfcon,bfopen)) #create bfi as a data frame as well describe(bfi.df) } #end of the stuff to be done within the with command 7 of 28 9/15/2016 1:16 PM

8 ) #end of the with command epi < person.data[c("epie","epis","epiimp","epilie","epineur") ] #form a new variable "epi" epi.df < data.frame(epi) #actually, more useful to treat this variable as a data frame bfi.df < data.frame(person.data[c(9,10,7,8,11)]) #create bfi as a data frame as well ls() #show the variables y < edit(person.data) #show the data.frame or matrix x in a text editor and save changes to y fix(person.data) #show the data.frame or matrix x in a text editor invisibible(edit(x)) #creates an edit window without also printing to console directly make changes. #Similar to the most basic spreadsheet. Very dangerous! head(x) #show the first few lines of a data.frame or matrix tail(x) #show the last few lines of a data.frame or matrix str(x) #show the structure of x 8 of 28 9/15/2016 1:16 PM

9 9 of 28 9/15/2016 1:16 PM

10 10 of 28 9/15/2016 1:16 PM

11 apply(epi,2,fivenum) #give the lowest, 25%, median, 75% and highest value (compare to summary) describe(epi.df) #use the describe function var n mean sd median trimmed mad min max range skew kurtosis se epie epis epiimp epilie epineur stem(person.data$bfneur) #stem and leaf diagram The decimal point is 1 digit(s) to the right of the round(cor(epi.df),2) #correlation matrix with values rounded to 2 decimals epie epis epiimp epilie epineur epie epis epiimp epilie epineur round(cor(epi.df,bfi.df),2) #cross correlations between the 5 EPI scales and the 5 BFI scales bfext bfneur bfagree bfcon bfopen epie epis epiimp epilie epineur corr.test(sat.act) > corr.test(epi.df) Call:corr.test(x = epi.df) Correlation matrix epie epis epiimp epilie epineur 11 of 28 9/15/2016 1:16 PM

12 epie epis epiimp epilie epineur Sample Size epie epis epiimp epilie epineur epie epis epiimp epilie epineur Probability values (Entries above the diagonal are adjusted for multiple tests.) epie epis epiimp epilie epineur epie epis epiimp epilie epineur of 28 9/15/2016 1:16 PM

13 13 of 28 9/15/2016 1:16 PM

14 datafilename=" project.org/r/datasets/r.appendix2.data" data.ex2=read.table(datafilename,header=t) #read the data into a table data.ex2 #show the data aov.ex2 = aov(alertness~gender*dosage,data=data.ex2) #do the analysis of variance summary(aov.ex2) #show the summary table print(model.tables(aov.ex2,"means"),digits=3) #report the means and the number of subjects/cell boxplot(alertness~dosage*gender,data=data.ex2) #graphical summary of means of the 4 cells attach(data.ex2) interaction.plot(dosage,gender,alertness) #another way to graph the means detach(data.ex2) #Run the analysis: datafilename=" project.org/r/datasets/r.appendix3.data" data.ex3=read.table(datafilename,header=t) #read the data into a table data.ex3 #show the data aov.ex3 = aov(recall~valence+error(subject/valence),data.ex3) summary(aov.ex3) print(model.tables(aov.ex3,"means"),digits=3) #report the means and the number of subjects/cell boxplot(recall~valence,data=data.ex3) #graphical output datafilename=" project.org/r/datasets/r.appendix4.data" data.ex4=read.table(datafilename,header=t) #read the data into a table data.ex4 #show the data aov.ex4=aov(recall~(task*valence)+error(subject/(task*valence)),data.ex4 ) summary(aov.ex4) print(model.tables(aov.ex4,"means"),digits=3) #report the means and the number of subjects/cell boxplot(recall~task*valence,data=data.ex4) #graphical summary of means of the 6 cells attach(data.ex4) interaction.plot(valence,task,recall) #another way to graph the interaction detach(data.ex4) datafilename=" project.org/r/datasets/r.appendix5.data" data.ex5=read.table(datafilename,header=t) #read the data into a table #data.ex5 #show the data aov.ex5 = aov(recall~(task*valence*gender*dosage)+error(subject/(task*valence))+ (Gender*Dosage),data.ex5) summary(aov.ex5) print(model.tables(aov.ex5,"means"),digits=3) #report the means and the number of subjects/cell boxplot(recall~task*valence*gender*dosage,data=data.ex5) #graphical summary of means of the 36 cells boxplot(recall~task*valence*dosage,data=data.ex5) #graphical summary of means of 18 cells datafilename="/users/bill/desktop/r.tutorial/datasets/recall1.data" recall.data=read.table(datafilename,header=true) recall.data #show the data 14 of 28 9/15/2016 1:16 PM

15 raw=recall.data[,1:8] #just trial data #First set some specific paremeters for the analysis this allows numcases=27 #How many subjects are there? numvariables=8 #How many repeated measures are there? numreplications=2 #How many replications/subject? numlevels1=2 #specify the number of levels for within subject variable 1 numlevels2=2 #specify the number of levels for within subject variable 2 stackedraw=stack(raw) #convert the data array into a vector #add the various coding variables for the conditions #make sure to check that this coding is correct recall.raw.df=data.frame(recall=stackedraw, subj=factor(rep(paste("subj", 1:numcases, sep=""), numvariables)), replication=factor(rep(rep(c("1","2"), c(numcases, numcases)), numvariables/numreplications)), time=factor(rep(rep(c("short", "long"), c(numcases*numreplications, numcases*numreplications)),numlevels1)), study=rep(c("d45", "d90"),c(numcases*numlevels1*numreplications, numcases*numlevels1*numreplications))) recall.aov= aov(recall.values ~ time * study + Error(subj/(time * study)), data=recall.raw.df) #do the ANOVA summary(recall.aov) #show the output print(model.tables(recall.aov,"means"),digits=3) #show the cell means for the anova table 15 of 28 9/15/2016 1:16 PM

16 16 of 28 9/15/2016 1:16 PM

17 #get the data datafilename=" project.org/r/datasets/extraversion.items.txt" #where are the data items=read.table(datafilename,header=true) #read the data attach(items) #make this the active path E1=q_262 q_1480 +q_819 q_1180 +q_ #find a five item extraversion scale #note that because the item responses ranged from 1 6, to reverse an item #we subtract it from the maximum response possible + the minimum. #Since there were two reversed items, this is the same as adding 14 E1.df = data.frame(q_262,q_1480,q_819,q_1180,q_1742 ) #put these items into a data frame summary(e1.df) round(cor(e1.df,use="pair"),2) round(cor(e1.df,e1,use="pair"),2) #give summary statistics for these items #correlate the 5 items, rounded off to 2 decimals, #use pairwise cases #show the item by scale correlations #define a function to find the alpha coefficient alpha.scale=function (x,y) #create a reusable function to find coefficient alpha #input to the function are a scale and a data.frame of the items in the scale { Vi=sum(diag(var(y,na.rm=TRUE))) #sum of item variance Vt=var(x,na.rm=TRUE) #total test variance n=dim(y)[2] #how many items are in the scale? (calculated dynamically) ((Vt Vi)/Vt)*(n/(n 1))} #alpha E.alpha=alpha.scale(E1,E1.df) #find the alpha for the scale E1 made up of the 5 items in E1.df detach(items) #take them out of the search path summary(e1.df) #give summary statistics for these items q_262 q_1480 q_819 q_1180 q_1742 Min. :1.00 Min. :0.000 Min. :0.000 Min. :0.000 Min. : st Qu.:2.00 1st Qu.: st Qu.: st Qu.: st Qu.:3.750 Median :3.00 Median :3.000 Median :5.000 Median :4.000 Median :5.000 Mean :3.07 Mean :2.885 Mean :4.565 Mean :3.295 Mean : rd Qu.:4.00 3rd Qu.: rd Qu.: rd Qu.: rd Qu.:6.000 Max. :6.00 Max. :6.000 Max. :6.000 Max. :6.000 Max. : of 28 9/15/2016 1:16 PM

18 18 of 28 9/15/2016 1:16 PM

19 19 of 28 9/15/2016 1:16 PM

20 20 of 28 9/15/2016 1:16 PM

21 21 of 28 9/15/2016 1:16 PM

22 22 of 28 9/15/2016 1:16 PM pnorm(1,mean=0,sd=1)

23 [1] pnorm(1) #default values of mean=0, sd=1 are used) pnorm(1,1,10) #parameters may be passed if in default order or by name samplesize=1000 size.r=.6 theta=rnorm(samplesize,0,1) #generate some random normal deviates e1=rnorm(samplesize,0,1) #generate errors for x e2=rnorm(samplesize,0,1) #generate errors for y weight=sqrt(size.r) #weight as a function of correlation x=weight*theta+e1*sqrt(1 size.r) #combine true score (theta) with error y=weight*theta+e2*sqrt(1 size.r) cor(x,y) #correlate the resulting pair df=data.frame(cbind(theta,e1,e2,x,y)) #form a data frame to hold all of the elements round(cor(df),2) #show the correlational structure pairs.panels(df) #plot the correlational structure (assumes psych package) library(mvtnorm) samplesize=1000 size.r=.6 sigmamatrix < matrix( c(1,sqrt(size.r),sqrt(size.r),sqrt(size.r),1,size.r, sqrt(size.r),size.r,1),ncol=3) xy < rmvnorm(samplesize,sigma=sigmamatrix) round(cor(xy),2) pairs.panels(xy) #assumes the psych package 23 of 28 9/15/2016 1:16 PM

24 24 of 28 9/15/2016 1:16 PM

25 25 of 28 9/15/2016 1:16 PM

26 26 of 28 9/15/2016 1:16 PM

27 hist() #histogram plot() plot(x,y,xlim=range( 1,1),ylim=range( 1,1),main=title) par(mfrow=c(1,1)) #change the graph window back to one figure symb=c(19,25,3,23) colors=c("black","red","green","blue") charact=c("s","t","n","h") plot(x,y,pch=symb[group],col=colors[group],bg=colors[condit],cex=1.5,main="main title") points(mpa,mna,pch=symb[condit],cex=4.5,col=colors[condit],bg=colors[condit]) curve() abline(a,b) abline(a, b, untf = FALSE,...) abline(h=, untf = FALSE,...) abline(v=, untf = FALSE,...) abline(coef=, untf = FALSE,...) abline(reg=, untf = FALSE,...) identify() plot(eatar,eanta,xlim=range( 1,1),ylim=range( 1,1),main=title) identify(eatar,eanta,labels=labels(energysr[,1]) ) #dynamically puts names on the plots locate() pairs() #SPLOM (scatter plot Matrix) matplot () #ordinate is row of the matrix biplot () #factor loadings and factor scores on same graph coplot(x~y z) #x by y conditioned on z symb=c(19,25,3,23) #choose some nice plotting symbols colors=c("black","red","green","blue") #choose some nice colors barplot() interaction.plot () #simple bar plot #shows means for an ANOVA design plot(degreedays,therms) #show the data points by(heating,location,function(x) abline(lm(therms~degreedays,data=x))) #show the best fitting regression for each group x= recordplot() #save the current plot device output in the object x replayplot(x) #replot object x dev.control #various control functions for printing/saving graphic files pnorm(1,mean=0,sd=1) [1] pnorm(1) #default values of mean=0, sd=1 are used) pnorm(1,1,10) #parameters may be passed if in default order or by name samplesize=1000 size.r=.6 theta=rnorm(samplesize,0,1) #generate some random normal deviates e1=rnorm(samplesize,0,1) #generate errors for x e2=rnorm(samplesize,0,1) #generate errors for y weight=sqrt(size.r) #weight as a function of correlation x=weight*theta+e1*sqrt(1 size.r) #combine true score (theta) with error y=weight*theta+e2*sqrt(1 size.r) cor(x,y) #correlate the resulting pair df=data.frame(cbind(theta,e1,e2,x,y)) #form a data frame to hold all of the elements round(cor(df),2) #show the correlational structure pairs.panels(df) #plot the correlational structure (assumes psych package) library(mvtnorm) samplesize=1000 size.r=.6 sigmamatrix < matrix( c(1,sqrt(size.r),sqrt(size.r),sqrt(size.r),1,size.r, sqrt(size.r),size.r,1),ncol=3) xy < rmvnorm(samplesize,sigma=sigmamatrix) round(cor(xy),2) 27 of 28 9/15/2016 1:16 PM

28 28 of 28 9/15/2016 1:16 PM pairs.panels(xy) #assumes the psych package